ObjectVideo deploys video analytic software to Port of Texas City

Published 19 September 2006

Deal follows on successes with Madrid, Spain and HSARPA, among others; software is able to detect violations of predetermined behavioral rules, allowing fewer employees to cover more ground; video analytics a new but growing field; stronger algorithms will take the industry into the future

Reston, Virginia-based ObjectVideo chalked up another success this week with the announcement that it had succesfully deployed (and will now test) its video analytic software on forty cameras surrounding the Port of Texas City. The location is a tricky one, with the U.S. Texas City Harbor on its eastern side, several chemical plants and oil refineries on its northern and western sides, and an industrial ship canal on its southern side. Detecting people, vehicles, and watercraft — and correctly distinguishing friend from foe — is difficult, requiring security personnel to set up event-specific, pre-defined rules. When an object violates a rule, the software alerts operations personnel in real time by phone, pager, email or an alert console. This allows those personnel to effectively monitor dozens of cameras simultaneously.

The Port of Texas City was no doubt influenced by earlier ObjectVideo successes. We reported on a number of them, including a deal with Madrid, Spain’s rail system inspired by recent bombings linked to abandoned backpacks. Company officials believe its software would have caught the terrorists leaving their backpacks behind on the train — suspicious behavior indeed. ObjectVideo has also developed new scanning camera technology for border applications for the Homeland Security Advanced Research Project Agency (HSARPA). All told, ObjectVideo technology is already deployed in more than eighty U.S. border locations and has contracts with the Department of Defense, the U.S. Air Force and Navy, the Department of Energy, and more than thirty commercial airports and seaports.

-read more in this news release


Trend: Intelligent video analytics come into their own

Not so long ago it was not easy to find a camera which could detect motion. Detecting motion, and other improvements in video technology, have given rise to a new term of art — video analytics. Not only has the technology improved, but there are now many more companies offering variants of it. Clara Conti, the CEO of Reston, Virginia-based IPIX says: “When I started, there were only two companies doing video analytics; Now there are 30 companies doing this or at least claiming to do this.”

Video analytics has moved beyond mere motion detection. Today the technology uses complicated algorithms to recognize different things — a person, a car, a fire, or anything else which a programmer can write an algorithm to recognize. This, however, is only the first step: Once an object has been defined, today’s sophisticated video systems allow users to set rules for action: If an object (a person, a vehicle) has crossed a predetermined line, for example, an alarm is sent to a security post and officers are dispatched to the scene. The systems can send alerts to pagers, emails, cell phones, or any other predefined device.

The direction of the technology: Develop complex algorithms which will recognize suspicious behavior of individuals in public places; identify suspicious objects, etc. Also, as is the case with WiFi access points, a debate is beginning on whether to embed the system’s intelligence at the edge (that is the camera itself), or in a central server.